Classify Documents in HCP according to given information

Dear community,

I am pretty new to Hana and I want to leverage its capabilities for some sort of Text Mining.

I have two tables, one with keywords linked to a class (Table "SpecialWords") and another table with a text (Table "Documents").

What I am thinking of is something like this.

Table SpecialWords

Table Documents

What I now want to create is a table which indicates what kind of text is stored in table "Documents" (e.g. Financial or IT). So add a Column to Documents which says "Financial" or "IT".

I already tried to use the AFM with the Naive Bayes Classifier but couldn't get any result out of it.

Does anyone had a similar problem and managed to solve it or has a recommendation for a good way to solve such problems?

I've also heard in an OpenSAP-course about Text Mining and the KNN-Classifier but couldn't find any technical approach to applying it.

Thanks in advance guys!